Video Quality Measurement Using Quantiser Step Size
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for measuring video quality after encoding and decoding introduce distortion, requiring time-consuming viewer opinions or inconvenient reference to the original sequence, especially when using variable quantiser step sizes and differential coding.
Innovation Solution
A no-reference, decoder-based quality assessment tool that generates a subjective quality estimate using a combination of sequence-averaged quantiser step size and regionally weighted contrast measures, applicable to video signals encoded with H.262 and H.264 standards, without needing the original signal.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If Mean Opinion Score (MOS) measurement is used to assess video quality, then measurement accuracy is improved, but measurement time increases significantly
Solution Approach 1:
The patent creates an objective quality metric that copies the essential characteristics of subjective MOS measurement by analyzing encoding parameters (quantizer step sizes, macroblock types, motion vector precision) and spatial complexity features. This objective copy reproduces MOS results without requiring actual human viewers, thus resolving the time consumption issue while maintaining measurement accuracy.
Solution Approach 2:
The patent replaces the mechanical human viewing and rating system with an automated computational system that processes encoding parameters and image statistics. This substitution eliminates the time-consuming human element while preserving the ability to measure video quality, directly addressing the contradiction between accuracy and time efficiency.
2Measurement precision
If reference to the original sequence is used to obtain degradation information, then measurement accuracy is improved, but ease of operation deteriorates
Solution Approach 1:
The patent extracts quality assessment capability from the original sequence itself by analyzing encoding parameters and decoded image statistics. This extraction eliminates the need to access or compare with the original sequence, making the method convenient for assessing archived or transmitted video where the original is unavailable, while maintaining accurate degradation measurement through encoding-parameter-based analysis.
Solution Approach 2:
The patent introduces encoding parameters (quantizer step sizes, macroblock types, motion vector precision) and spatial complexity measures as intermediaries between the encoded video and quality assessment. These intermediaries provide sufficient information to evaluate quality without requiring direct access to the original sequence, thus improving ease of operation while preserving measurement accuracy.
3Productivity
If variable quantiser step size is used in encoding, then bandwidth efficiency is improved, but video quality uniformity deteriorates
Solution Approach 1:
The patent applies local quality analysis by examining quantizer step sizes and spatial complexity on a per-macroblock basis rather than globally. This localized approach captures the non-uniform quality distribution caused by variable quantizer step sizes, allowing accurate quality assessment that reflects both bandwidth efficiency gains and local quality variations, thus resolving the contradiction between bandwidth efficiency and quality consistency.
Data Source
Figure 1
Figure 1a
Figure 1b
AI summary
This invention is concerned with a video quality measurement method, in particular where the video signal having: an original form; an encoded form in which video signal has been encoded using a compression algorithm utilising a variable quantiser step size such that the encoded signal includes a quantiser step size parameter; and, a decoded form in which the encoded video signal has been at least in part reconverted to the original form. The method comprises the steps of: a) generating a first quality measure which is a function of said quantiser step size parameter; b) generating a masking measure and c) combining the first and second measures. The masking measure is which is a function of the spatial complexity of parts of the frames represented by the video signal in the decoded form, these parts being selected by generating a second measure which is a function of the prediction residual and identifying one or more regions of the picture for which the second measure exceeds a threshold.